It has been shown that fluctuation of human heartbeat intervals [heart rate variability (HRV)] reflects variations in cardiac autonomic nervous system activity. The present study was designed to investigate whether the acute exposure to moderate levels of simulated altitude and the resultant hypoxia could modify HRV during exercise. Seven healthy men completed one resting measurement in the upright sitting position and two submaximal steady-state cycle ergometer exercises at intensities equivalent to 25 and 50% of their estimated maximal work rate. Experiments were conducted in random order at altitude equivalents of 500, 1,500, 2,500, and 3,500 m within 2 h of exposure to that altitude. Beat-to-beat HRV was measured continuously during the tests. HRV data were analyzed by "coarse-graining spectral analysis" (Y. Yamamoto and R.L. Hughson, Physica 68D: 250-264, 1993) to break down their total power (PT) into harmonic and nonharmonic (fractal) components. The harmonic component was further divided into low (0.0- to 0.15-Hz; PL)- and high (> 0.15-Hz; PH)-frequency components, and the indicators of relative sympathetic (SNS) and parasympathetic (PNS) nervous system activities were calculated by PL/PH and PH/PT, respectively. The fractal component was used to calculate the spectral exponent (beta) to evaluate the overall "irregularity" of HRV. The effects of exercise intensity (increase in heart rate, SNS indicator, and beta and decrease in PNS indicator) were significant (P < 0.05) at all altitude levels. The altitude effects (increase in heart rate and SNS indicator and decrease in PNS indicator) were found only during exercise at 3,500 m (P < 0.05). There was no significant effect of altitude on beta (P > 0.05). These data indicate that acute effects of altitude exposure on HRV were found 1) during exercise at moderate altitude (> 2,500 m) and 2) mainly for the harmonic components of HRV.